5  Buildings

The enclaves of “old Bombay” are worlds apart from the tumultuous jumble of urban form (Chapter 3) of the neighbourhood I grew up in. The ornate buildings in the tree-lined neighbourhoods of Colaba or Parsi Colony in Dadar are like an elegant elderly lady, sagging and covered in liver-spots, but her charm and beauty still take your breath away.

Like any city in India (and perhaps more so), Mumbai is a city of contradictions and contrast. The razzle dazzle of Bollywood make it the city of dreams but the vast droves of rural migrant workers living in slums also make it the city of inequality.

India loves to contradict itself on everything and anything. […] Every stereotype you know of India is completely true. At the same time the stereotypes are being broken all the time. Love is forbidden but romance is blossoming. India is poor, but it is a land of luxury. It has blazing hot summers and snows in the winter. It has the most barren deserts and the lushest jungles…

5.1 Buildings anchor the city

Buildings anchor urban form and their spread and style also dictate the aesthetics. This is even more the case for urban areas with a large number of informal dwellings (independent or as part of a slum cluster). Within slums, buildings dictate movement more than street networks via narrow alleys (“gullies” in Mumbai parlance) rather than formally managed and designed streets.

Despite a substantial slum population (last calculated ~42% in 2011 [2]), the ingenuity of the people living in these areas exceeds imagination. Facilitated by the cheek-by-jowl proximity of labour and business, slums are economic powerhouses where no task is too small. For example, the recycling of hair waste to wigs and hair extensions is a multi-million dollar industry that employs a vast segment of entrepreneurial individuals and businesses from ragpickers (collecting discarded hair from the streets) to local waste collection depots and wig artisans [3].

If you look over the sprawl extending out from Indian cities, you will see that there are a lot of informal developments that are not slums. The more affluent of these settlements differ in terms of their density and the materials from which they are built, and they tend to be made up of single family homes with private space. Yet the form of the settlement, the layout of streets and spaces, and the relationship of homes to each other is not so different from the poorest settlements made out of corrugated sheet and tarpaulin.

  • Rudlin and Hemani. Climax city (2019) [4]

Shivaji Nagar.

Mankhurd.

5.2 Getting building data

The R package osmdata can be used to query for building polygons in a small area. A helper function is used to simplify the generation of a bounding box around a central point. The quality of OSM in New Zealand is very high since LINZ manages regular uploads through their process of deriving outlines from satellite imagery.

# load packages to get data
library(dplyr)
library(sf)
library(osmdata)
library(osmplotr)
source("R/centroid_bbox.R")

akl_bbox <- get_centroid_bounding_box(
  c("lat" = -36.85494696219434, "lng" = 174.76040843528585),
  distance = 1000,
  dist.unit = "m") %>% 
  get_bbox()

akl_bldgs_osm <- opq(akl_bbox) %>%
  add_osm_feature("building") %>% 
  osmdata_sf()

akl_bldgs_osm_poly <- bind_rows(
  akl_bldgs_osm$osm_polygons,
  akl_bldgs_osm$osm_multipolygons
) %>% 
  st_make_valid()
# load plotting library
library(tmap)
tmap_options(check.and.fix = TRUE)

# plot buildings on map
tmap_mode("plot")
tm_shape(akl_bldgs_osm_poly) + 
  tm_polygons()

Figure 5.1: Map of buildings in Central Auckland from OpenStreetMap (OSSM).

Querying building data for a large area from OSM is not performant. For larger scale analyses, LINZ has building outlines available to download for the whole country or for a cropped section.

# get buildings from LINZ export
# https://data.linz.govt.nz/layer/101290-nz-building-outlines/
akl_bldgs <- st_read("./extdata/nz-building-outlines.shp")
# plot buildings around Auckland Harbour
tmap_mode("plot")
tm_shape(akl_bldgs) + 
  tm_polygons()

Figure 5.2: Map of buildings around Auckland Harbour from LINZ.

Building data quality and availability is an issue for analyses of developing nations - where the lower resources and capacity of local municipalities cannot keep up with the pace of informal development. Building shapes derived from satellite imagery help to address this issue (somewhat). Both Google and Microsoft provide global coverage of building footprints. The projects are in their nascence but it’s likely that they will only get better over time. TODO: Global ML is not that great at sub-dividing the tightly clustered urban form of slums. Satellite imagery of this area to show the high density of very small buildings.

The highest quality building available currently is the EUCUCCO dataset [5]. This dataset allows for detaileed analysis of urban form as it includes attributes like building type (use), height and date of construction. With such a rich dataset, the evolution and type of urban form can be traced through time. It’s especially useful as a comparsion to theories in urban economics.

Some of the building attributes available in EUBUCCO. Example of Paris reproduced from [5].

5.3 Figure-ground diagrams

Figure-ground diagrams highlight built and unbuilt space. Only buildings are included in the conventional approach resulting in a stark mosaic of black and white that allows for a bird’s eye analysis of urban form and structures. However, the diagram can be adjusted by adding green space, water bodies and streets with a tentative width corresponding to its hierarchy for more insights about the urban tissue.

By showing buildings in black and removing most other detail, they [the figure-ground] reveal the truth about a place. Like X-rays, they are calibrated to see through the tissue of detail on a map or aerial photo, revealing the underlying structure of an urban area. In doing so, the plans reveal the shattering and decay of many industrial towns and cities as well as the inner beauty and structure of ‘healthy’ urban places. You can look at the figure ground plan of a city and ‘read’ the place: you can generally tell where the centre is and trace the transept to its suburbs, you can follow the line of the most important commercial streets, identify which districts are thriving and which are declining.

  • Rudlin and Hemani. Climax city (2019) [4]

Figure-ground diagrams capture a square mile section of an urban area [6] - a rounded 2km\(^2\) in metric terms.

Given the highly heterogeneous urban mosaic of megacities like Mumbai, smaller areas are more meaningful as they represent a consistently designed space. For this reason, the figure-ground diagrams in Figure 5.3 have a 500m bounding box around a central coordinate.

The famous rectangular blocks of New York City (NYC) are a stark contrast to the more sinuous blocks of British (local and colonial) urbanism. Secondary or tertiary roads surround all blocks in NYC while blocks in Islington and Dadar Parsi Colony, two residential neighbourhoods of British design, are connected by residential roads with arterial connections much further away.

Figure 5.3: Modified figure-ground diagrams (with street widths) for neighbourhoods in Mumbai, London and New York City. Decreasing width in order of OSM classification - primary, secondary, tertiary, residential.

5.4 Urban morphology

5.5 Learning resources